The Future of Work Podcast

Good afternoon, and welcome to the ILC lunchtime Conversation

brought to you by the ILO's Future of Work podcast.

We're coming to you today from the 113th International Labour

Conference at the Palais des Nations here in Geneva.

And today we're talking about a very important topic:

AI and the world of work and the transformations

that the digital transformation and AI is having on the world of work.

Now, this is an area that the ILO is leading on

when it comes to both research and policy recommendations.

And it's an area that makes a lot of people worried and anxious.

A lot of people wonder, will AI replace me as a worker?

Will I be able to learn the skills

I need to learn in order to work in this era of digitalization?

And if I do learn those skills, will they become antiquated

because the digital transformation is moving so fast?

And what do I need to do to make AI work for me as a worker?

Well, to discuss all of this we have with us

today two fantastic researchers at the ILO.

Here on my right is Pawel Gmyrek, Senior Researcher

in the Effective Labour Institutions Unit in the Department of Research.

Welcome.

And also with us is Manal Azzi, the ILO Team Lead

on Occupational Safety and Health Policy and Systems.

Welcome, both.

Pawel, I'd like to start with you.

You recently put out an interesting report where you found

that one out of four jobs globally will be transformed,

transformed, not replaced or disappeared.

That sounds like positive news, is it?

Well, it very

much depends on the job and the country you're in.

So let me first clarify that with our work.

And the report you mentioned is focussed on generative

AI, because that's what a lot of us are asking about.

So these chatbots that we all know by now and what they can do

with existing jobs and

okay, and basically

what we do is we examine what people do at work

by looking at the tasks of different occupations.

So if we think about the tasks of an ILC delegate,

you have a number of tasks you have to accomplish in terms

of preparing the trip, preparing the positions, the briefings,

participating in meetings, hearing discussions, making statements,

going back, reporting, developing policy positions, and so on.

And if we take each of those tasks, we basically try to understand

to what extent could a technology like, let's say ChatGPT

or another one of those AI-driven chatbots,

or technology with similar capacities,

to what extent they could perform these tasks

that you have to perform coming here, but also that workers

have to perform in occupations around the world.

So we try to go over all types of occupations and run the same exercise.

And once we've done it, we try to pull this data together and

and link it to statistics and understand how many such jobs exist around the world.

And what we did is we organize them

into jobs that are exposed to AI and not exposed in our view.

And the good news is, okay, 75% of all employment

today, in our view, doesn't have that much of exposure to these tools.

So that number, that's three quarters of people who go to work every day.

And 25% of those who go to work

daily will come into some type of interaction

with this technology based on different levels of possible automation of the tasks.

And we organized this group into four subgroups, which we call the gradients.

So if you are in gradient 1 or 2 you probably have a few tasks

that this technology can affect and possibly automate with time

but where human

skill is still very much needed.

And if you go to gradients three and four, you find yourself in occupations

where a large number of tasks today could potentially be automated with time.

But that doesn't mean that such an occupation

has disappeared also it might mean a more rapid evolution.

So I think maybe what's

worth highlighting is in that last group,

the most exposed occupations,

we also see something very interesting.

We have these occupations, which we already knew about,

these administrative occupations or clerical jobs,

which have more of tasks exposed to the risks of automation.

And we knew about it before,

but now we see these increasing abilities of this technology

to also automate tasks

in highly digitized, specialized occupations like programmers,

software developers, or financial advisors, analysts,

insurance advisors.

So a lot of very technical, digital jobs where these tools are now

acquiring the possibility to automate a wider range of tasks.

But as I said, it's not the end of work.

It's more of a transformation that we are talking about.

And in some occupations that transformation

will be more dynamic, in others it will be at a different pace.

Absolutely.

And we will come back

to what is needed to ensure that the transformation is a positive one

in a minute. I want to turn over to Manal.

Manal, you recently issued a paper in which you outlined the ways in which

artificial intelligence,

digitization, robots, which is my favourite, are actually

being deployed to protect workers, both in terms of their health and their safety.

That also sounds like good news.

Can you tell us a little bit more about that?

Thank you Zeina.

And building on what Pawel was saying, we did launch this year, this report,

we have extra copies later on revolutionizing safety and health.

So in our branch on occupational safety and health, we looked at how different

technologies, automation,

digitalization have improved conditions of safety and health.

We know from different research also that work is changing

even in assembly, in manual handling, in hazardous environments.

All this automation

and digitalization is changing these work tasks. For safety and health,

in most part, it's been quite positive.

Robots and automation

have taken over some of the very hazardous and dangerous jobs.

We know that they can work at high temperatures.

They can work with the high repetitive movements

and jobs in production processes.

So they take over these repetitive movements.

But we also know that the development of smart wearables and smart devices

and sensors has allowed us to not only detect hazards early on, but also

predict hazards and become more proactive in managing safety and health.

And there are so many examples, Zeina on that, including,

personal protective equipment like helmets that now have toxic gas sensors.

We have ear defenders that now can detect noise in advance

and also alert workers to stop work when it becomes very hazardous,

not to mention robots that are even, you know, taking on work in surgery,

saving a lot of time,

removing a bit of the workload for healthcare workers and also

very efficient and allowing workers to do more stimulating tasks.

I think that's the positive twist to it, is that

we don't have to do repetitive and meaningless, less stimulating tasks.

We now can move on to more meaningful, exchanges with each other,

even as we deal with patients, as we deal with different jobs.

The other point we looked at is algorithmic management of work.

And if we look at that positively, we see that

it allows us to understand the preferences.

We're collecting data.

We know the scheduling preferences of staff and other

gamification processes that we can use to reward staff as they work.

And the last area we looked at in the report is the changing work arrangements.

While this, you know, can be seen as negative in some ways,

it does bring a lot of positives

with platform work, teleworking that allows a more inclusive market

where people with caregiving challenges

or with disabilities can actually access the market by working

in these teleworking, schedules and platform economy.

But of course, Zeina,

there's a problem with all this and it doesn't come without new risks.

And so the report explores that, you know, we can't over rely on automation

and hazard detection through these digitalized, and AI-powered systems.

The human needs to stay at the centre.

The human needs to think of the longer-term risks, not just the short-term risks

that they're being faced. And also robots and machines,

they can fail.

And software failures have been reported as one of the key issues.

And not to mention, last but not least, the monitoring.

So yes, you can monitor to prevent hazards,

but sometimes over monitoring can lead to people

being judged on their productivity or being nudged to be more productive.

So those are the different risks, and there's so much more

that are displayed in the report on that end.

Yeah, absolutely.

And I think you're going to exactly the heart of the next part

that I'd like to explore with you guys,

which are the risks and the potential pitfalls.

There's been a lot of work, some of which is from the ILO

and also other organizations about

the digital divide and how that plays out.

The inequalities, the invisible workers behind

AI and, and the data gathering and so on.

So there are a number of issues that one must be aware

of when we're talking about, digitalization in AI.

And maybe I can come back to you,

Pawel, in your research, what are the key issues that have emerged

that we as the ILO, but also as just international community need to be aware

of when we're looking at digitization in general and AI in particular.

Well, there is of course a lot,

but I think in terms of the direct impact on occupations, what is really important

is to understand the local context, to make it context specific.

Because as we saw in the recent report,

there's a huge variety of potential impacts when you look across

regions, income levels or the types of occupations.

And when we make it context specific, when we link it to existing occupations,

we can understand a little bit better that map.

And who are the people at risk of maybe going through

some employment transitions or maybe even losing jobs?

How many such people there are?

And who are the occupations?

Who are the people in those jobs who might actually benefit from it,

but might be limited from accessing those benefits today

because of these existing constraints?

And I think, you know, mapping this out is step number one.

Number two is basically making the tools and mechanism of social dialogue

work for policy development so that these policies can be developed

in a more targeted way based on that knowledge.

Because as we know, administrations always operate in

constraints in terms of financial constraints

and how much focus can be put on different areas.

So if we know exactly which people we are talking about,

you can design much more targeted policies to respond to that.

And of course, there is that digital divide

issue that you mentioned that is very present in many countries.

So your story or your experience

as a worker will really depend a lot on where you are sitting geographically

and what occupation you're in. If you're in this highest exposure gradient.

I mentioned, gradient four, in a high-income economy,

it might be a rough time, might be a time when there's a lot of transition

happening.

And already the transition is something that we know has a big impact in terms

of also mental health and emotional,

you know, impact on workers, even if we're not talking about unemployment.

And you could also have such effects

if you are in these lower gradients of exposure,

a gradient one and two in a low-income country,

you are more likely to actually be in a position where you might miss out

on opportunities that this technology could offer to increase your productivity

to be more competitive.

So that's why making it context specific

and bringing it to mechanism of dialogue can help

design much better local policies,

and also to think about it as a process, not something that gets solved once,

but it's a process that we have entered that might even be accelerating,

so having these mechanisms in place will help

sort of navigate through that process as things develop.

And this technology

is also developing really fast is what we also showing in our report.

Indeed, you mentioned policy. Policymaking

is very much at the centre of this whole process.

So now if we take, turn our attention

a little bit to the ILO, how would you say the ILO is contributing

to the global conversation about AI and the ongoing digital transformation

in a way that benefits employers,

governments, as well as workers?

Would you?

Yeah, please.

So we are in the International Labour Conference, and it's very important

to note that today we have the first discussion

on promoting decent work in the platform economy.

So that really falls as one of the biggest endeavors

the ILO is doing today. It will be discussed, these two weeks and the next year

in the conference to come up with conclusions and potentially

a standard, hopefully, where you all have discussed and agreed

on how do we ensure decent work in the platform economy?

One group of, workers, actually, we need to understand that

when we talk about AI, when we talk about digitalization is the whole supply chain.

So you have workers extracting minerals to ensure that you're able to use this

and create the tools you need

for video systems that are AI-powered and other technologies. And these

people working sometimes in very poor working conditions or poor safety

and health conditions, we need to make sure we understand.

And the whole process of transport until waste and waste management

of these electronic products, sometimes can expose workers

to high levels of hazardous substances and chemicals.

So we've looked

at the whole supply chain of AI, and that also requires policy responses.

And from the national level, and the way the ILO supports different countries

is to revise regulations so that we take into account

potential accidents that can happen between robots and human interaction.

The right to disconnect.

You're all attached to your internet and your computers.

And so when is it the right to disconnect?

And how can that be regulated? In addition to workplace-level risk assessments

for the safety and health, not only to skill

workers on new technologies and the technicalities

of the new technologies, but also how to stay safe by using the technologies

and to explain to them why the technologies are being introduced

and to allow opportunities to skill, reskill, upskill people

as we use these technologies and keep these people in the workforce.

So there are so many levels that workers need to be informed

at the workplace and trained and be a true partner in the application

and use of AI-powered systems and other technologies at the workplace.

Absolutely.

And the ILO is very much at the heart of that discussion.

Pawel?

Yeah.

So coming from the Research department, I think for the start,

I mean, our contribution is to try to provide knowledge

about the existing processes and the way we see them advance.

We try to provide it

as basically an input to different mechanism of social dialogue

that can be used on that basis in a more informed way.

And we also try to create a centralized repository of that knowledge.

So that's why we launched the ILO Observatory.

And I invite you to follow us on LinkedIn and to visit our page,

because that's where we post

all the information that we gather on the basis of research,

And we think about it as a platform for collaboration

also with the field structure and with research is also ongoing in the field.

Because at the end of the day,

we don't only provide research

from our side and knowledge, but we have so much to learn from constituents

where you observe these things first hand.

And pulling this together into that centralized knowledge platform

would help a lot

also across regional learning and sharing of experiences.

And we see a lot happening at the national level.

And we think there is more that we can pull together into

these type of exchanges.

So one such opportunity will be next week here on Thursday

12th June, we will have a side event just around here.

And we would invite you to participate in that because that will be

a great opportunity to share your experiences with digitalization on

AI and different themes that are part of that event with others.

And of course,

there are

a lot of other collaboration opportunities,

also with our colleagues in the field.

So we basically will try to pull it together on the Observatory.

But a lot of this will happen locally and we learn from that as well.

And we offer a space for this kind of

joint learning as this process advances.

Great.

And the AI Observatory is a great tool.

I do encourage you to look it up online.

The ILO AI Observatory.

Interesting discussion, but I'm afraid that's all the time we have for

today's episode.

Thank you all for listening.

We hope you'll join us again

for more conversations on key issues shaping the world of work.

And don't forget to follow us on social media.

You can find us on X: "@ILO", as well as Facebook: "@International Labour

Organization", LinkedIn: "International Labour Organization",

Instagram "ilo_org"

and YouTube: "ILO TV".

Thank you very much for being with us today.

Thank you to our speakers.

And until the next time, take care from all of us at the ILO.